On quasi-isospectrality of potentials and Riemannian manifolds

Imagine you are in a dark room with two different musical instruments. You can't see them, but you can listen to the notes they play when you strike them. In mathematics, this is called the isospectral problem: If two instruments produce the exact same set of notes (frequencies), are they actually the same instrument, or just different ones that happen to sound identical?

This paper, written by Clara L. Aldana and Camilo P´erez, explores a slightly more relaxed version of this mystery. They ask: What if the instruments are almost identical in sound, differing by just one note? Can we still tell if they are the same, or does that single missing note change everything?

Here is a breakdown of their findings using simple analogies:

1. The Setup: The Drum and the Shape

The classic version of this problem was famously asked by mathematician Mark Kac: "Can you hear the shape of a drum?"

The Drum: Imagine a drumhead (a surface). When you hit it, it vibrates at specific frequencies. These frequencies are the "spectrum."
The Shape: The physical shape and material of the drum determine those frequencies.
The Mystery: If two drums sound exactly the same, do they have the same shape? Sometimes yes, sometimes no.

2. The New Twist: "Quasi-Isospectrality"

The authors introduce a new concept called quasi-isospectrality.

The Analogy: Imagine two pianos. Piano A plays a perfect scale. Piano B plays the exact same scale, except for one specific key that is slightly out of tune (or perhaps a different key is missing).
The Question: If two objects (like a drum or a mathematical "potential" which acts like the tension on the drum) are almost identical in their sound, are they actually the same object?

3. The Magic Trick: The "Darboux" Spell

To create these "almost identical" instruments, the authors use a mathematical technique called the Darboux transformation (or the "double commutation method").

The Metaphor: Think of this as a magical spell. You take a drum, pick one specific note, and use the spell to slightly shift that note's pitch.
The Catch: Usually, doing this breaks the drum (creating mathematical "singularities" or tears in the fabric). However, the authors show that if you cast the spell twice in a specific way, the tears heal, and you end up with a brand new, perfectly smooth drum that sounds almost exactly like the original, just with that one note tweaked.

4. The Big Discovery: Odd vs. Even Dimensions

This is the paper's most surprising finding. The authors discovered that the answer to "Are they the same?" depends entirely on the dimension of the space the object lives in.

Odd Dimensions (The "Rigid" World):
- Imagine a 1D string or a 3D drum.
- The Result: If two objects in an odd-dimensional space are "quasi-isospectral" (differ by one note), they are actually isospectral (identical).
- The Analogy: It's like trying to build a 3D sculpture out of clay where you change the color of just one tiny speck. In this specific mathematical universe, the laws of physics (or math) are so strict that if the sound is almost the same, the shape must be exactly the same. You cannot have that one different note without changing the whole shape.
Even Dimensions (The "Flexible" World):
- Imagine a 2D drumhead or a 4D hyper-sphere.
- The Result: Here, you can have two different shapes that are quasi-isospectral. They can differ by that one note and still be different objects.
- The Analogy: In this world, you can tweak that one note and the shape can adjust slightly to accommodate it without becoming a completely different object. However, the authors prove that even here, the differences are tightly controlled and bounded.

5. Why Does This Matter?

The paper also looks at compactness.

The Metaphor: Imagine a box full of all possible drums that sound exactly the same. Is this box finite and manageable, or is it infinite and chaotic?
The Finding: The authors show that even with this "quasi" (almost) definition, the set of possible shapes remains "compact" (manageable and finite in a mathematical sense) for low-dimensional spaces. This means that even if you allow for that one slightly different note, you can't wander off into infinite, unmanageable variations of shapes.

Summary

In everyday terms, this paper is a detective story about sound and shape.

The Crime: Two objects sound 99% the same, differing by only one note.
The Investigation: The authors use a "double spell" (Darboux transformation) to create these near-matches.
The Verdict:
- If the object exists in an odd dimension (like our 3D world), the suspect is guilty of being identical to the original. The "almost same" sound proves they are the same object.
- If the object exists in an even dimension, they might be different, but they are still very closely related and follow strict rules.

The paper essentially draws a line in the sand: In the mathematical universe, the dimension of space dictates whether a tiny difference in sound implies a tiny difference in shape, or if it forces the entire shape to be identical.

Here is a detailed technical summary of the paper "On quasi-isospectrality of potentials and Riemannian manifolds" by Clara L. Aldana and Camilo P´erez.

1. Problem Statement

The paper addresses the inverse spectral problem, specifically investigating the rigidity and compactness properties of operators when the strict condition of isospectrality (identical spectra including multiplicities) is relaxed to quasi-isospectrality.

Context: In classical spectral geometry, two operators are isospectral if their spectra are identical. A famous question posed by Mark Kac ("Can one hear the shape of a drum?") asks if the spectrum uniquely determines the geometry (or potential). While counterexamples exist (isospectral but non-isometric manifolds), rigidity results often hold under specific constraints (e.g., low dimensions, specific boundary conditions).
The Gap: The authors introduce and analyze quasi-isospectrality, a weaker notion where two operators share the same spectrum except for one eigenvalue, which is allowed to vary within a prescribed small interval. The central questions are:
1. Do quasi-isospectral manifolds or potentials retain the rigidity properties of isospectral ones?
2. Can we construct such operators systematically?
3. Do compactness results for isospectral sets extend to quasi-isospectral sets?

2. Methodology

The authors employ a combination of spectral theory, inverse spectral methods, and heat kernel asymptotics.

Heat Trace Analysis: The core analytical tool is the asymptotic expansion of the heat trace, $\text{Tr}(e^{-t\Delta})$ $Tr (e^{- t Δ})$ , as $t \to 0^+$ $t \to 0^{+}$ .
- For closed manifolds of dimension $n$ : $\text{Tr}(e^{-t\Delta}) \sim t^{-n/2} \sum a_j t^j$ .
- For manifolds with boundary: The expansion involves half-integer powers ( $t^{j/2}$ ).
- The coefficients $a_j$ (heat invariants) are geometric integrals involving curvature and the potential.
Darboux Transformation (Double Commutation Method): To construct quasi-isospectral potentials on intervals, the authors utilize the Bilotta-Morassi-Turco (BMT) method. This involves applying Darboux's Lemma twice:
1. The first application introduces a singularity (removing an eigenvalue).
2. The second application removes the singularity, resulting in a regular potential with a modified spectrum (one eigenvalue shifted).
Spectral Zeta Functions: The authors use relative spectral zeta functions to relate the regularized determinants of quasi-isospectral operators.

3. Key Contributions and Results

A. Construction of Quasi-Isospectral Potentials (Sections 3–5)

The paper provides a systematic construction of quasi-isospectral Sturm-Liouville operators on a finite interval $[0,1]$ with Dirichlet boundary conditions.

Method: By applying the double commutation method to a base potential $q$ , they construct a new potential $p$ such that $\text{Spec}(T_p)$ differs from $\text{Spec}(T_q)$ by exactly one eigenvalue $\lambda_n \to \lambda_n + t$ .
Boundary Conditions: The authors analyze how boundary conditions are preserved or altered. They show that applying Darboux's lemma to eigenfunctions of the Dirichlet problem yields a new potential that also satisfies Dirichlet conditions, provided specific limits are taken.
Examples:
- They construct a family of even quasi-isospectral potentials to the zero potential, demonstrating that evenness is preserved under this specific construction.
- They exhibit examples where operators with different boundary conditions (e.g., Robin vs. Dirichlet) can be isospectral or quasi-isospectral.

B. Rigidity of Quasi-Isospectral Manifolds (Section 7)

This is the paper's primary theoretical breakthrough regarding Riemannian manifolds.

Main Theorem (Theorem 7.2): Let $(M, g)$ $(M, g)$ be a closed Riemannian manifold of dimension $n$ $n$ . If two metrics $h_1$ $h_{1}$ and $h_2$ $h_{2}$ are $(k, \epsilon)$ $(k, ϵ)$ -quasi-isospectral (differing only in the $k$ $k$ -th eigenvalue within $\epsilon$ $ϵ$ ), then:
1. Odd Dimensions ( $n$ is odd): $h_1$ and $h_2$ are strictly isospectral. The heat trace expansion contains only integer powers of $t$ for the difference, while the spectral difference introduces a term that cannot be matched by integer powers unless the shift is zero. Thus, no non-trivial quasi-isospectral deformation exists in odd dimensions.
2. Even Dimensions ( $n$ is even): The metrics are not necessarily isospectral. However, their heat invariants $a_j$ coincide for $j < 1 + n/2$ . For higher indices, the difference $|a_j(h_1) - a_j(h_2)|$ is bounded by a term proportional to $\epsilon$ .
Regularized Determinant: The authors show that for quasi-isospectral metrics, the ratio of their regularized determinants is exactly the ratio of the differing eigenvalues: $\frac{\det(\Delta_g)}{\det(\Delta_h)} = \frac{\lambda_k(g)}{\lambda_k(h)}$ .

C. Quasi-Isospectral Potentials on Manifolds (Section 8)

The authors extend the rigidity results to Schrödinger operators ( $\Delta + V$ ) on closed manifolds.

Odd Dimensions: If $n$ is odd, any sequence of quasi-isospectral potentials is actually isospectral.
Even Dimensions: The heat invariants are preserved up to a certain order, with higher-order invariants differing by $O(\epsilon)$ .
Compactness: Building on classical results by Brüning and Donnelly, the authors prove that the set of quasi-isospectral potentials remains compact in the $C^\infty$ topology for dimensions $n \leq 3$ (general case) and $n \leq 9$ (for non-negative potentials). This is achieved by showing that the heat invariants provide uniform Sobolev bounds, even with the slight perturbation in the spectrum.

D. One-Dimensional Case (Section 8, Corollary 8.5)

For Sturm-Liouville operators on $[0,1]$ with Dirichlet conditions:

Since $n=1$ is odd, quasi-isospectral potentials are isospectral.
Consequently, the mean value of the potential $\int_0^1 q(x) dx$ and other spectral invariants must be identical for any two quasi-isospectral potentials.

4. Significance and Implications

Dimensional Dichotomy: The paper establishes a fundamental distinction between odd and even dimensions in spectral geometry. In odd dimensions, the spectral data is so rigid that even a single eigenvalue perturbation (quasi-isospectrality) forces the operators to be identical. This is a strong rigidity result that generalizes previous isospectral theorems.
Generalization of Compactness: The authors successfully extend the compactness of isospectral sets (a cornerstone of inverse spectral theory) to the quasi-isospectral setting. This implies that small spectral perturbations do not lead to "exploding" families of potentials; the set remains well-behaved in Sobolev spaces.
Constructive Theory: By detailing the BMT method and providing explicit examples, the paper bridges the gap between abstract spectral theory and concrete construction of counterexamples or specific spectral data sets.
Heat Invariant Sensitivity: The analysis clarifies exactly which geometric quantities (heat invariants) are sensitive to a single eigenvalue shift. In even dimensions, only the higher-order invariants (related to higher-order curvature derivatives) are affected, while lower-order geometric data (volume, Euler characteristic, scalar curvature integrals) remain invariant.

Conclusion

Aldana and P´erez demonstrate that quasi-isospectrality is a very strong condition, particularly in odd dimensions where it collapses entirely into isospectrality. In even dimensions, it acts as a controlled perturbation that preserves the lower-order geometric structure of the manifold or potential. These results deepen the understanding of the relationship between spectral data and geometric structure, offering new tools for inverse problems in mathematical physics.